Automated 3D and 4D organ delineation for radiation therapy planning in the pelvic area

The determination of the treatment parameters in radiation therapy requires the segmentation of the patient anatomy. This procedure is usually performed by manual contouring of 2D slices, which may require several hours. The burden is considerably increased in the context of intensity modulated radiation therapy (IMRT) and 4D adaptive radiotherapy, which aims at compensating the significant anatomical changes during the course of treatment based on additional imagery. The development of fast and robust automated segmentation tools is crucial for these novel treatment methods to succeed. The purpose of this paper is to automate organ contouring of 3D CT data and time series of 3D CT in radiation therapy planning (RTP). We present an automated model-based concept for organ delineation, based on adaptation of interactively positioned 3D deformable surface models to the boundaries of the anatomical structures of interest in 3D CT. In order to make the concept clinically feasible, we introduce interactive tools for efficient correction in problematic areas where the automated model adaptation may fail. A feasibility study with 3D CT datasets of 40 clinical cases was done for the risk organs (bladder, rectum, and femoral heads) of the pelvic area. In several cases minor user interaction was required, nevertheless a significant reduction of time compared to manual organ contouring was achieved (minutes vs. hours). The results of the validation study demonstrate that the presented model-based approach is accurate (1-1.7mm mean error) for the tested anatomical structures. We investigate the extension to 4D by adaptating 3D deformable models to each 3D image dataset of the time series, and show illustrative results on a clinical case with 16 datasets. Quantitative analysis shows comparable results to the 3D interactive method for the femur and the bladder.